Explore recent advancements in functional optimization through this insightful 59-minute conference talk by Nicolas Le Roux from Microsoft, MILA, McGill University, and Université de Montréal. Delve into the argument for analyzing optimization problems in function space rather than parameter space, as traditionally done in classical continuous optimization. Discover new theoretical results and innovative practical algorithms applied to standard problems in reinforcement learning and supervised learning. Gain valuable insights into this evolving field of study and its potential implications for various machine learning applications.
Overview
Syllabus
Recent Advances in Functional Optimization, Nicolas Le Roux
Taught by
GERAD Research Center
Related Courses
-
Combining Reinforcement Learning and Constraint Programming for Combinatorial Optimization
-
Learning to Optimize High Dimensional Optimization Problems
-
Faster Saddle-Point Optimization for Solving Large-Scale Markov Decision Processes
-
Offline Reinforcement Learning and Model-Based Optimization
-
Hitting the High-Dimensional Notes - An ODE for SGD Learning Dynamics in High-Dimensions
